Nodes/ComfyUI_MusicTools/Music - Stereo Enhance
ComfyUI Node

Music - Stereo Enhance

Wider Stereo Without the Phase Nightmare

By jeankassio·Created 9 months ago·Updated a day ago· 154
Music - Stereo Enhance
  • audio
  • AUDIO
intensity0.5

Wide stereo is the cheapest way to make a track feel "produced" - and the easiest way to break it. Music - Stereo Enhance from ComfyUI Music Tools does the standard mid/side trick: it takes the mid (center) and side (left-right difference) components of your audio, leaves the mid alone, and turns up the sides. One input, one output, one knob.

The one knob

intensity - 0 to 2, default 0.5. That's the entire control surface. 0 is a no-op, 0.5 is a sensible default widening, and past ~1.0 you're in "check it on headphones, then check it in mono" territory, because every stereo widener has a phase cost and this one is no exception.

The node also handles mono input gracefully: a mono track gets safely duplicated to stereo before widening. If you've ever fed a mono vocal into a stereo widener and watched it collapse into a hollow mess, you know why that matters - the pack explicitly guards against the classic mono-in-stereo-out trap.

Where it fits

Stereo enhancement is a finishing move. It belongs after EQ and compression, around the same spot you'd put a limiter - it's shaping the presentation of the mix, not fixing it. A typical finishing tail:

Load Audio -> Music Fix -> Stereo Enhance -> LUFS Normalizer -> Save Audio

For generated audio it pulls double duty: AI music tends to come out of the model narrow and center-heavy, so a touch of width is often exactly what makes a render sound less synthetic. That's the pack's actual intended use case - it was built with ACE-Step and Suno-style output in mind, per the pack's own launch framing.

The honest take

At default intensity it's a genuinely nice, safe widener. Push it hard and you'll hear the mid/side artifacts - the classic symptom is the lead vocal (mid-heavy) sounding thin while the hats (side-heavy) swim around it. The mono-compat check is the one habit that saves you: if it sounds great on headphones but hollow on a phone speaker, you widened too far. Start at 0.3, not 1.5.

There's also a version of this baked into the Music Fix profiles (stereo shaping is part of the finishing chain, minus bass widening, which the pack deliberately avoids - widening the low end is how you get the wobbly phase mess). So if you're already using Music Fix, the standalone node is mostly for when you want width without the rest of the finish.

Install

Everything in this pack installs the same way:

cd ComfyUI/custom_nodes
git clone https://github.com/jeankassio/ComfyUI_MusicTools.git
cd ComfyUI_MusicTools
python -m pip install -r requirements.txt

Restart ComfyUI, find it under music. No models, no GPU, runs near real-time - it's a handful of NumPy operations on the side channel.

Categorymusic

Inputs (2)

NameTypeDefaultDescription
audioAUDIO
intensityFLOAT0.50–2

Outputs (1)

NameTypeDescription
AUDIOAUDIO